Skylight

Skylight picked $79 over $99 because customers felt $99 was unfair. Tests showed $99 a year would earn slightly more than $79, but customer interviews put $99 near 'disgust' or 'not worth it yet' territory, so Skylight shipped $79. Going from $39 to $79 cost minimal retention or attach and was described as transformative to the ARR curve.

No retention, attach or revenue figures are given for the increase, and the claim that $99 would have earned more is not backed with numbers. The team pairs interviews with quant tests because people sometimes say one thing and do another, and can revisit the price as more value is added. "We look in the whites of customers' eyes and we try to figure out what feels relatively reasonable," says CEO Michael Segal.

Sub Club▶ 51:12

Test subscription price and hardware discount together, not separately. Skylight varied the subscription price and the device discount at once across many variants, then modeled the combined effect on hardware conversion, margin per unit and subscription attach in a spreadsheet before choosing $79.

"The relationship between those two changes the conversion in hardware and the attach to software."

Sub Club▶ 50:28

Grandfather existing subscribers and honor the old price for hesitant buyers. Skylight did not announce the increase and moved existing customers to a 'legacy plus' plan at $39. Anyone who had been considering a purchase for months and contacted support also got the old price, effectively indefinitely.

The team had math showing that squeezing legacy customers was worth a couple million dollars and skipped it to protect goodwill. "Be really nice to the existing customers, build for the next 10 million customers and don't try to squeeze every penny out."

Sub Club▶ 53:58

A device used every day makes a subscription easy to justify. The calendar is the 'heartbeat of your home', so it supports a subscription better than a photo frame sitting at grandma's house. Skylight launched in 2015 and added the subscription in 2018, when it had a couple million in revenue, two people and an NPS 'off the charts'.

Co-founder Michael Segal, a former Bessemer VC, held that consumer hardware without a subscription builds no enterprise value, with rare exceptions. Paying a fee on top of a $300 to $600 device still feels uncomfortable to buyers, so the price has to be seen as fair again every year.

Sub Club

Keep the job the customer bought the device for free. On the frame, sending photos to grandma is free and video is the upsell. On the calendar, events, tasks and chores are included while recipes, meals and an AI engine are paid.

CPO Mark Ungerer's test is whether the person came, spent the money and got what they wanted; if not, they feel 'tricked and duped'. Marketing loudly advertising paid features like the AI engine risks turning excitement into frustration after purchase.

Sub Club▶ 11:09

Launch new features behind the paywall rather than moving them there later. Meal planning launched as a paid feature, and many such launches are not A/B tested at all; the gut check is whether buyers still got what they paid $300 for. Moving a feature out of the paywall is easy, moving one in is 'very painful'.

Paid features should be quick to understand, not discovered as a locked setting mid-configuration. Rebundling and remerchandising have not changed attach rate much, since excited buyers want the whole thing.

Sub Club▶ 16:01

Treat customer anger at a paywall as a cost even when numbers rise. Some paywalled items lifted metrics but drew begrudging or angry reactions in reviews and interviews, and Skylight reversed course on them.

Segal calls the subscription 'not a get-rich-quick scheme' and says disgust is the emotion to avoid when building a multi-billion dollar brand.

Sub Club

Paywall choices made early can sour as the customer base matures. The photo screensaver went behind the $39 paywall with no rigor, on the logic of buying for the calendar and paying to also get a frame. Eight years later more customers say it should be free, and the team is debating making it fully or partly free.

Early buyers were eager enthusiasts; later buyers are more skeptical. Kids love the photos and happy owners tell friends, so fully paywalling it may be shortsighted.

Sub Club

Don't make subscribers tally which features are premium. Skylight does not label premium features inside the product for subscribers. Renewal rates are high both for heavy bundle users and for those who barely use it.

Hardware buyers go all in after months of research and want to know the device is valuable and used daily, not run a per-feature value calculation.

Sub Club▶ 22:46

On its own site Skylight bundles the subscription with a hardware discount. On skylight.com the subscription is turned on automatically with a free month and the hardware is discounted if you sign up with it. Users can opt out but most don't, so attach is much higher than on Amazon or in stores.

"We give a discount on the hardware if you sign up with the subscription and that is a super successful way to get subscribers."

Sub Club▶ 42:19

Retail lowers attach but works as a stamp of quality and a moat. Skylight sells through its own site, Amazon, Costco, Best Buy and internationally. Being everywhere has driven higher overall growth despite lower subscription attach in retail.

Retailers usually take only one player per category, which makes shelf presence defensible. Lower attach makes post-purchase education more important. "Retail is the biggest influencer of all in our world."

Sub Club▶ 40:38

Routing the device QR code to Apple in-app purchase doubled conversion. Non-subscribed users scanning the device's QR code used to land on a web checkout. Switching to Apple in-app purchase gave over a 100% lift in conversion on a smaller group of users.

Skylight frames this as growth versus profit: web checkout avoids the 30% fee, but once lost conversion from extra clicks is counted the saving is closer to 20%. With no plan to sell or chase EBITDA, it chooses growth.

Sub Club▶ 56:31

A free trial added nothing when buyers already subscribe at checkout. Skylight tested a free trial with and without a credit card, and neither variant increased conversion. Trial takers were roughly the same people who would have subscribed anyway, and over half of calendar buyers already subscribe.

Buyers research for months and decide before opening the app. With no growth PM and about 5% of the roadmap, Skylight could not run the follow-up tests it needed: "the first test is going to fail and you need to run like eight more." It is now hiring its first subscription growth PM.

Sub Club▶ 57:27

Get families to set up more features in the first days. Within days or weeks of setup, families settle into habits and Skylight's ability to get them to try new things 'drops to zero'. Customers who set up with more features use them more.

Customers often say they know the device does much more but never explore it. Skylight is exploring conversational, AI-driven onboarding to fit more aha moments into the first 10 minutes, such as building a chore chart from kids' ages.

Sub Club▶ 43:17

Marketing below about 40 NPS gives a false negative. The calendar shipped in 2018 and a 2021 to 2022 marketing push failed because the product wasn't ready. The team nearly killed it, rebuilt the product, and it took off in fall 2022.

Below roughly 40 NPS you can't tell whether marketing or product is failing. "If your product isn't like a solid 40 NPS, I think marketing effort and dollars are kind of a waste of time."

Sub Club▶ 39:14

Skylight tracks the share of customers who call it life-changing. About 20% of customers polled say the product is life-changing, and the goal is to build features that push that toward 50%.

The hypothesis is that each of these customers brings in dozens of new ones by word of mouth.

Sub Club▶ 41:07

Selling hardware up front lets every order pay back on day zero. Every order in Skylight's history has paid back on day zero, so it never needed VC money to fund marketing. It is bootstrapped, profitable and in about 2 million households.

For a software subscription, the equivalent is pushing a roughly $60 to $70 annual payment up front to fix CPA and payback. "Do whatever you humanly can to pay back on day zero."

Sub Club▶ 29:47

Keeping some hardware margin gives room to absorb shocks. Subscriptions are about 10 to 20% of revenue but half of profit, with hardware making the other half. That margin let Skylight hold prices through tariffs and the memory price spike by giving up hardware profit this year.

A hardware price cut also remains available later to help reach 20 million households.

Sub Club▶ 28:20

Neither hardware nor software is a lasting moat. Competitors cloned Skylight's UI and sold it through hundreds of Chinese OEMs on Amazon. Short term the defense is the best product and speed; long term it has to be network effects, partnerships, retail relationships or a developer platform.

Whatever the advantage, it must be authentic to the product. Segal also advises entrepreneurs against launching hardware with a subscription, since few break out: about twenty brands now versus about three five years ago.

Sub Club▶ 35:45